Remove date from logger

[pascalduez]

So [18:26:48] instead of superfluous [2015-01-22 18:26:48]

Thoughts ?

[valeriangalliat]

The date was not needed indeed, neither is the time IMO.

[pascalduez]

The date was not needed indeed, neither is the time IMO.

Time kind of make sense in a debug scenario, where one could merely check execution time... ?
But usually SassDoc performs in less than a second.

[pascalduez]

Maybe the normal (non debug) verbose mode should rather looks like something in this vein ?

[18:45:09] SassDoc task started
> Item "global-test" refers to "mixin-global-test" from type "mixin" but this item doesn't exist.
> Item "global-test" refers to "variable-global-test" from type "variable" but this item doesn't exist.
> Item "global-test" refers to "placeholder-global-test" from type "placeholder" but this item doesn't exist.
> Folder "./test/data" successfully parsed.
> Folder "sassdoc" successfully refreshed.
> Theme "default" successfully rendered.
> Process over. Everything okay!
[18:45:09] SassDoc task ended

[KittyGiraudel]

Maybe the normal (non debug) verbose mode should rather looks like something in this vein ?

Looks nice.

[valeriangalliat]

where one could merely check execution time

time sassdoc ...

[pascalduez]

So shall we go for the fancier formatting ? Or just keep it as is ?

» Item "global-test" refers to "mixin-global-test" from type "mixin" but this item doesn't exist.
» Item "global-test" refers to "variable-global-test" from type "variable" but this item doesn't exist.
» Item "global-test" refers to "placeholder-global-test" from type "placeholder" but this item doesn't exist.
» Folder "./test/data" successfully parsed.
» Folder "sassdoc" successfully refreshed.
» Theme "default" successfully rendered.
» Process over. Everything okay!
✔ SassDoc task ended in 200ms

[KittyGiraudel]

Could we have an octopus icon? :x

[KittyGiraudel]

Boom: http://www.fileformat.info/info/unicode/char/1f419/index.htm. 🐙

[pascalduez]

Current status:

Still room for improvements.

[FWeinb]

I like it ✔

[KittyGiraudel]

👍

[valeriangalliat]

Looks good!

[pascalduez]

[valeriangalliat]

Errors have a red » too?

[pascalduez]

Those are just 'log' messages, I kept the [WARNING] and [ERROR] ones we already had.
Should we streamline all messages ?

[FWeinb]

Making it red on an error and yellow on a warning sound awesome

[valeriangalliat]

Should we streamline all messages ?

I think so, but I don't know how. It's nice to differenciate heads-up from success with yellow/green but if we add this scheme for warning messages, they'd be yellow too…

Though "but this item doesn't exist" messages are kinda warnings to me.

[valeriangalliat]

I think we need to change the logger semantics (and function names):

function	color
log (same as info)	default
success	green
warning	yellow
error	red

[pascalduez]

Looks good. So I would also put the whole line colored for warning and error.
Should we keep the [WARNING] and [ERROR] prepends ?

» [WARNING] You'd better watch out bro, something's broken.

» [ERROR] You're screwed anyway.

I give it a try so we can check how it looks.

[valeriangalliat]

Visually, with colors, we don't need [WARNING] and [ERROR], though it's really helpful when logs are saved to text files to search through them. it's the only way on non-colored terminals to differenciate informational messages from warnings and errors.

An alternative would be to replace the » with I W or E for respectively info, warning, error (in the corresponding color).

I don't think the whole line needs to be colored, only the » or "semantic leter" is enough.

[FWeinb]

I would include [WARNING] and [ERROR] and only color the chevron.

[pascalduez]

Okay for only having the prepended thing colored.

Although I like the I W or E I don't think it's straightforward enough for end users ?

[pascalduez]

I would include [WARNING] and [ERROR] and only color the chevron.

👍 let's try

[valeriangalliat]

I don't know, I recall seeing this in unit test suites. Having the [] is okay too.

[FWeinb]

Let's look at the output of popular unit test frameworks

[pascalduez]

I don't know, I recall seeing this in unit test suites. Having the [] is okay too.

What I meant is it's maybe a bit too nerdy :-) We have devsigner™ between SassDoc users.

[pascalduez]

Stacktrace truncated from screenshot

[FWeinb]

This looks awesome! But just name it SassDoc completed after xxxms

[valeriangalliat]

Perfect!

[pascalduez]

And the logger code has been considerably lightened/cleaned.

Also now that annotations use warn instead of log, the warnings appears in the tests.
But I guess it even better, warnings should be considered, not ignored.

[pascalduez]

Mr @hugogiraudel the last word maybe ?

[KittyGiraudel]

You guys rock. LGTM.

Le Sam 24 Jan 2015 12:34, Pascal Duez notifications@github.com a écrit :

Mr @hugogiraudel https://github.com/HugoGiraudel the last word maybe ?

—
Reply to this email directly or view it on GitHub
#333 (comment).

[pascalduez]

Thanks team for the great and constructive feedback !

[KittyGiraudel]

Tried it today, looks gorgeous. Love it.

[pascalduez]

And its now fully tested / testable.

[coveralls]

Changes Unknown when pulling bc8957b on simplify-logger into * on develop*.

[pascalduez]

@coveralls U drunk?

[valeriangalliat]